Los Angeles Classification of Esophagitis using Image Processing Techniques

Abstract

Esophagitis is a condition of inflammation of the esophageal mucosa, which is also called as acid reflux disease. The cause maybe be due to slackness of the lower esophageal sphintcer which allows acidic contents of the food from stomach to esophagus. Esophagitis is detected by observing the esophagus by video endoscopy of the Upper Gastro-Intestinal tract. The classification of esophagitis is done by analyzing the images captured during the process of endoscopy. Classification of Esophagitis has many standards , with each standard having its plus and minus. The Los Angeles(LA) Classification deals with precise measurement of the mucosal breaks, for an image processing system to measure the mucosal breaks the position of the camera is to be known. We attempt to classify the Esophagitis using LA Classification without the camera position information using low level image features and classification is performed using a neural network classifier. The results of the classifier are compared with inter and intra observer variability studies.